Power grid enterprises collect a large number of data resources in the process of production management, including management documents, technical standards, project application documents, experimental reports, technical reports, etc. The location and retrieval speed of related text file data determine the efficiency of problem solving. Some related work has built a semantic search system in power field based on knowledge map [1], but only some preliminary ideas and methods are given. This paper introduces a power text entity recognition technology based on deep transfer learning Bert bilstm CRF, which can get better accuracy without a large amount of manual intervention. The experimental results show that the accuracy of the model in system, technology, equipment and other professional fields is greatly improved compared with the traditional annotation method.
Higher Accuracy Entity Recognition of Power Grid through Bert-BiLSTM-CRF
2021-10-20
1164958 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Named entity recognition of railway dispatching command based on BiLSTM-CRF
British Library Conference Proceedings | 2023
|MOVING ENTITY RECOGNITION METHOD, MOVING ENTITY RECOGNITION SYSTEM, AND ENTITY RECOGNITION EQUIPMENT
Europäisches Patentamt | 2016
|